netdev - Re: RMNET QMAP data aggregation with size greater than 16384

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 6 Aug 2021 16:08:45 +0200
From:   Aleksander Morgado <aleksander@...ksander.es>
To:     Subash Abhinov Kasiviswanathan <subashab@...eaurora.org>
Cc:     Bjørn Mork <bjorn@...k.no>,
        Daniele Palmas <dnlplm@...il.com>,
        Network Development <netdev@...r.kernel.org>,
        Sean Tranchetti <stranche@...eaurora.org>
Subject: Re: RMNET QMAP data aggregation with size greater than 16384

Hey Subash,

> > I may be mistaken then in how this should be setup when using rmnet.
> > For the qmi_wwan case using add_mux/del_mux (Daniele correct me if
> > wrong!), we do need to configure the MTU of the master interface to be
> > equal to the aggregation data size reported via QMI WDA before
> > creating any mux link; see
> > http://paldan.altervista.org/linux-qmap-qmi_wwan-multiple-pdn-setup/
> >
> > I ended up doing the same here for the rmnet case; but if it's not
> > needed I can definitely change that. I do recall that I originally had
> > left the master MTU untouched in the rmnet case and users had issues,
> > and increasing it to the aggregation size solved that; I assume that's
> > because the MTU should have been increased to accommodate the extra
> > MAP header as you said. How much more size does it need on top of the
> > 1500 bytes?
>
> You need to use an additional 4 bytes for MAPv1 and 8 bytes for
> MAPv4/v5.
>

I tried with a SIMCOM 7600E, with data aggregation enabled with QMAPv1:

$ sudo qmicli -d /dev/cdc-wdm0 -p --wda-get-data-format
[/dev/cdc-wdm0] Successfully got data format
                   QoS flow header: no
               Link layer protocol: 'raw-ip'
  Uplink data aggregation protocol: 'qmap'
Downlink data aggregation protocol: 'qmap'
                     NDP signature: '0'
Downlink data aggregation max datagrams: '10'
Downlink data aggregation max size: '4096'

As you suggested, the MTU of the new muxed interface is set to 1500
and the MTU of the master interface to only 4 more bytes (1504):

# ip link
8: wwp0s20f0u8u4i5: <POINTOPOINT,UP,LOWER_UP> mtu 1504 qdisc fq_codel
state UNKNOWN mode DEFAULT group default qlen 1000
    link/none
9: qmapmux0.0@...0s20f0u8u4i5: <UP,LOWER_UP> mtu 1500 qdisc fq_codel
state UNKNOWN mode DEFAULT group default qlen 1000
    link/[519]

Under this scenario, the downlink is completely broken (speedtest
0.39Mbps), while the uplink seems to work (speedtest 13Mbps).

If I use the logic I had before, associating the downlink data
aggregation max size reported by the module to the MTU of the master
interface, same as I had to do when using qmi_wwan add_mux/del_mux,
then it works properly:

# ip link
14: wwp0s20f0u8u4i5: <POINTOPOINT,UP,LOWER_UP> mtu 4096 qdisc fq_codel
state UNKNOWN mode DEFAULT group default qlen 1000
    link/none
15: qmapmux0.0@...0s20f0u8u4i5: <UP,LOWER_UP> mtu 1500 qdisc fq_codel
state UNKNOWN mode DEFAULT group default qlen 1000
    link/[519]

Downlink is now 26Mbps and uplink still 13Mbps.

Is there something I'm doing wrong? Or do we really need to do the
same thing as in qmi_wwan add_mux/del_mux; i.e. configuring the master
interface MTU to be the same as the downlink max aggregation data size
so that we change the rx_urb_size?

-- 
Aleksander
https://aleksander.es